The statins are a widely-used class of drugs that lower LDL (low density lipoprotein) cholesterol by inhibiting the enzyme 3-hydroxy-3-methylglutaryl-coenzyme-A (HMG-CoA) reductase and thereby reducing the production of cholesterol by the liver. Large-scale randomised evidence shows that statin therapy reduces the incidence of heart attacks, strokes and revascularisations by about one fifth per 1 mmol/L LDL-cholesterol reduction.1 The benefits achieved with statin use appear to relate primarily to an individual's absolute risk of such events, and to the absolute size of the LDL cholesterol reduction. The additional benefits seen with more intensive statin therapy have resulted in a trend towards the use of higher doses of statin.
Rarely, statins can cause muscle pain or weakness with elevated blood levels of creatine kinase (i.e. myopathy) and, in a minority of cases, this may lead to muscle breakdown and myoglobin release into the circulation (i.e. rhabdomyolysis) with a risk of renal failure and death.2 The mechanisms by which statins cause myopathy remain unknown, but they appear to be related to statin concentrations in the blood. The incidence of myopathy is only about one per 10,000 patients per year with standard statin doses (such as 20-40 mg simvastatin daily),3 but the risk increases (possibly about ten-fold) with higher statin doses (such as 80 mg simvastatin daily4). It is also increased by concomitant use of certain drugs that interact to produce raised plasma statin levels. For example, gemfibrozil given concomitantly has been found to increase the area under the statin elimination curve (AUC) by 2-4 fold with several statins and to increase myopathy risk many-fold.5,6 Concomitant use of cyclosporine and itraconazole, and other drugs that inhibit the CYP3A4 enzyme, has also been shown to increase plasma statin exposure several-fold and has been linked with myopathy.7,8 These clear associations have led to warnings in the statin drug labels against the concomitant use of particular statin doses and certain other drugs (especially gemfibrozil, cyclosporine and itraconazole).2 
Such interactions are thought to occur when the statin and the concomitant drug share common metabolic pathways. The gemfibrozil interaction with statins has been postulated to be mediated via the UGT glucuronidation enzyme genes or via several CYP genes.5,7 Several statins (including lovastatin, simvastatin and atorvastatin) are mainly metabolised via the CYP3A4 enzyme, and it has been concluded that most of the clinically important drug-drug interactions that occur with these statins are attributable to the concurrent use of agents that are potent inhibitors or substrates of CYP3A4.7 Pravastatin is not metabolised via the CYP genes, but its plasma level may be influenced by genes involved in its elimination by transportation. Although rosuvastatin metabolism does not appear to depend on the CYP system, several drug interactions of clinical significance are known. For example, when rosuvastatin is combined with cyclosporine, the AUC of rosuvastatin increased 7-11 times and it has been suggested that cyclosporine inhibition of organic anion transport polypeptide C may decrease hepatic uptake of rosuvastatin9 
The effects of more than 20 genes on statin pharmacokinetics have been investigated.4, 10 For five of these genes (SLCO1B1, CYP3A5, CYP2C9, ABCG2, ABCC2), at least one small study has reported associations with plasma statin levels. The SLCO1B1 gene encodes the organic anion transport protein OATP1B1 that is known to affect the hepatic uptake and biliary excretion of various drugs. In vitro studies indicate that most statins and statin acids are substrates for the SLCO1B1 transporter,11 although it has been suggested that its contribution to hepatic uptake is lower for lipophilic statins (such as simvastatin and lovastatin) which are thought to be taken up chiefly through passive infusion.12 A literature search undertaken by the present inventors found that 14 separate reports of the impact of the SLCO1B1 gene on statin pharmacokinetics (mostly involving pravastatin or rosuvastatin) had been published. Not all of the studies yielded statistically significant results, and a combined analysis of them had not previously been performed but the typical impact on statin pharmacokinetics was much smaller than the several-fold increases produced by concomitant use of gemfibrozil or potent CYP3A4 inhibitors. Consequently, it was not clear whether such differences in statin plasma levels would be of much relevance to the risk of statin-related myopathy.
Some small studies had previously considered the direct relevance to possible statin-related muscle side-effects of various candidate genes, such as CYP3A4 which is involved in the metabolism of certain statins,13 genes involved in ubiquinone (coenzyme Q10) deficiency,14 and genes encoding organic anion transporting polypeptides (OATP).11 Associations for myopathy, myalgia or statin intolerance had been reported at “nominal” p<0.05 (i.e. before making allowance for the large number of candidate genes and SNPs that were considered) with six genes individually. Given their small size and multiple comparisons however, these small studies did not provide good a priori evidence for any genetic associations with statin-related myopathy. Moreover, apparent differences in myopathy rates in those studies may have been confounded by differences in statin dosages and concomitant use of other drugs.3 In particular, one study15 of 10 patients with myopathy and 26 controls reported an association between myopathy among patients taking pravastatin or atorvastatin and the SLCO1BI *15 haplotype (rs4149056 C allele and rs2306283 G allele) with a nominal p-value<0.01. This small study involved the exploration of associations with 152 SNPs in different genes (as well as some haplotype comparisons) and with three separate statins (as well as different combinations of those statins). The impact of this large number of multiple comparisons needs to be allowed for when interpreting the nominal p-values: since not all of the tests would have been independent, the effective number of independent tests was between 300 and 1000. Hence, application of the Bonferroni approach would involve multiplying each nominal p-value by at least 300, rendering a nominal p-value value of 0.01 completely non-significant.